DESCRIPTION: (Applicant's Abstract) Rigorous diagnostic systems are prerequisite to understanding the etiology, epidemiology, prevention, treatment, topography and long-term course of substance use disorders. Consistent with the principal recommendation of a recent NIDA Technical Review, this is a proposal to conduct enhanced data analysis in one of NlDA's already well published nosologic studies, the Rutgers Research Diagnostic Project (RDP), a multisite, regional study of more than 400 cases funded by NIDA since 1989. Two additional sources of data will be analyzed. One contains diagnostic data on 348 clinical and community cases aged 12-19; the second contains data on 50 clinical cases aged 60+. The research will be conducted by a cross-institutional group of nosologists at Rutgers University, the University of Pittsburgh, the University of Iowa, and Columbia University. They will conduct a total of 7 study series on the nosology of substance use disorders across a wide socioeconomic and age range, in pursuit of 4 technical and 3 theoretical goals. The specific technical goals include (a) at the criterion level, the rigorous application of classical psychometrics vs. techniques not yet applied to the nosology of substance use disorders based on item response theory; (b) at the composite algorithm level, use of other techniques new to substance use disorders receiver-operator characteristic (ROC and QROC) analysis - to set optimal thresholds for diagnostic positivity and to designate efficient symptoms as candidate gate criteria; (c) studies of the predictive validity of DSM-IV and ICD-10 diagnoses, and of optimal diagnostic thresholds, at 12-month outcomes, and (d) study of experimental and drug-customized algorithms. Specific theoretical goals include (e) work on the dimensional vs. categorical (taxonic) nature of, and construct validation of, current models of drug dependence using a variety of methods: factor analysis of cross-sectional data, symptom onset and staging of longitudinal data, and taxometric analysis by distribution shape, MAXCOV/HITMAX, and Golden's (1982) method; (f) work on a consensual heuristic for mild and prodromal illness categories (abuse or harmful use), categories which are necessary to advance addictions research but which have not been well formulated; and (g) research on the clinical topography of drug problems through study of risk, illness onset, course, and individual differences variables, particularly age, comorbidity, special population, and gender effects.